AcrIIA5 (Streptococcus phage D4276, UniProt A0A220GHA5) — curation notes

Provenance of this review

Deep-research providers were unavailable in this container (no API keys), so no
-deep-research-<provider>.md file was generated. The literature below was assembled
directly from the cached publications in publications/, PubMed MCP searches
(AcrIIA5, 15 hits total), and from two additional papers that I cached during this
review with just fetch-pmid (PMID:31722192, PMID:31775029).

The GONUTS/CACAO wiki pages that would record the students' own evidence statements
(https://gowiki.tamu.edu/wiki/index.php/PMID:28785032 and .../PMID:30046034) were
offline at the time of review ("This site is temporarily offline while we perform
upgrades"), so the CACAO rows below were assessed against the cited papers directly.

Gene symbol choice

UniProt gives this entry no gene name (GN ORFNames=D4276_028), so just fetch-gene
fell back to the accession for gene_symbol. I set gene_symbol: acrIIA5, following
the repository's own anti-CRISPR precedents genes/LISMN/acrIIA4 (gene_symbol: acrIIA4) and genes/NEIME/acrIIC1 (gene_symbol: acrIIC1), and matching the folder
name. acrIIA5 is the name used throughout the primary literature and is the name
UniProt itself carries in DE SubName: Full=AcrIIA5; the Acr nomenclature (type,
subtype, serial number) is the community-standard naming system for anti-CRISPRs, so
it is the closest thing to an authoritative symbol for a phage ORF of this kind.

What the protein is

• 140 aa, basic, monomeric: PMID:32544231.
• Novel fold with an N-terminal intrinsically disordered region (IDR):
  PMID:32544231 and PMID:32544231. NMR structure PDB 6LKF; UniProt FT REGION 1..23 /note="Disordered" (MobiDB-lite) agrees with the experimental IDR assignment.
• Encoded by a virulent (strictly lytic) phage, not a prophage — this matters for the
  cellular-component call below: PMID:28785032.

Discovery and in vivo anti-CRISPR activity (CACAO-cited papers)

PMID:28785032 (Hynes et al. 2017, Nat Microbiol) — the discovery paper, cached
ABSTRACT ONLY (full_text_available: false). Phage D4276 escaped CRISPR immunity far
more often than its relatives: PMID:28785032. Systematic cloning
located the responsible gene: PMID:28785032 The protein also
blocks the heterologous SpCas9 system: PMID:28785032 and PMID:28785032. The assay throughout is phage plaquing on CRISPR-immunized
hosts, which is exactly what the CACAO rows' ECO code says (ECO:0007040 = plaque assay
evidence used in manual assertion).

PMID:30046034 (Hynes et al. 2018, Nat Commun) — full text available. Principally the
AcrIIA6 discovery paper, but AcrIIA5_D4276 (the 140-aa protein of this entry) is itself
assayed side by side in Table 1 and Figs. 4–5, so this is not a "paper about another
gene": PMID:30046034, PMID:30046034, PMID:30046034, and PMID:30046034. Note the one
quantitative limit specific to this orthologue: PMID:30046034 — i.e. D4276 AcrIIA5 is
potent against St1Cas9 and SpCas9 but only weakly active against St3Cas9, whereas the
D1126 orthologue fully inhibits St3Cas9.

Breadth of Cas9 targets

PMID:31722192 and PMID:31722192. Garcia et al. characterised the
D4276 protein itself, so this is direct evidence for this accession. Independently,
PMID:31775029 (abstract only in cache).

Molecular mechanism

Inhibition is of nuclease activity, not of target-DNA engagement:
PMID:31775029 and PMID:31775029.

In vitro with purified protein: PMID:32544231
(Fig. 3A). Specificity: PMID:32544231.

The binding target is the guide-loaded Cas9 ribonucleoprotein, not free RNA and not
apo-Cas9: PMID:32544231 (Fig. 4A, EMSA; corroborated by NMR line-broadening,
Fig. 4E), and PMID:32544231. Summary statement:
PMID:32544231.

Determinants: PMID:32544231; PMID:32544231; and
PMID:32544231 (the β3–β4 loop, where the in vivo loss-of-function residues Asp50/Arg62/Asp74/
Lys88/His66/Asn70/His73 of Garcia et al. map).

A contested finding worth flagging

Garcia et al. reported sgRNA cleavage in vivo and inferred an RNA-directed mechanism:
PMID:31722192,
reinforced by PMID:31722192. An et al. then tested AcrIIA5 for intrinsic
ribonuclease activity and found none: PMID:32544231, concluding PMID:32544231. So the observation of sgRNA cleavage stands, but the inference that
AcrIIA5 itself cleaves (or must engage) RNA is overturned for the catalytic reading and
disputed for the "mechanism involves RNA interaction" reading. Garcia et al. also note
the practical reason the mechanism was hard to pin down:
PMID:31722192, whereas An et al. succeeded by using an MBP tag and 500 mM NaCl.

Consequences for the annotation review

1. GO:0098672 symbiont-mediated suppression of host CRISPR-cas system (3 CACAO rows,
   IDA) is correct and maximally specific; the evidence is direct and on this protein.
   The one defect is the qualifier on the 2025 row:
   acts_upstream_of_or_within_positive_effect understates a process the protein itself
   executes; involved_in (used by the other two rows) is right. Recorded as a
   source_defects entry (WRONG_QUALIFIER) rather than as a REMOVE.
2. GO:0003723 RNA binding (DisProt, IPI, WITH RNAcentral sgRNA) is the wrong
   granularity and the wrong entity: the assayed interaction is with the Cas9–sgRNA
   ribonucleoprotein, and the paper that the IPI cites is the one that showed AcrIIA5 has
   no RNase activity and binds only the preformed complex. MODIFY to GO:0043021
   ribonucleoprotein complex binding (the term the repo's acrIIA4 review uses for the
   same Acr–Cas9/sgRNA relationship).
3. GO:0140678 molecular function inhibitor activity (DisProt, IMP) is true but
   uninformative when the target enzyme class is known. MODIFY to GO:0140721 nuclease
   inhibitor activity (an is_a descendant via GO:0004857), as used for acrIIC1 in this
   repo. I stopped short of GO:0060703 deoxyribonuclease inhibitor activity: AcrIIA5's
   substrate-level target is the Cas9 RNP rather than a classical DNase, and its reported
   effect is on the RuvC domain of an RNA-guided enzyme, so the parent term is the
   defensible stopping point.
4. GO:0030430 host cell cytoplasm added as NEW/IC. This entry's taxon is the phage
   (NCBITaxon:2006928), and the phage is virulent rather than a resident prophage, so —
   unlike the genes/LISMN/acrIIA4 case, which is filed under the Listeria taxon and
   therefore uses GO:0005737 — the compartment AcrIIA5 occupies belongs to a separate host
   cell. No localisation experiment exists; the location is a curator inference from the
   target (cytoplasmic Cas9 RNP) plus in-cell inhibition, hence IC.

Open questions

• Which surface of Cas9 does the IDR contact, and does it occlude RuvC directly or act
  allosterically? An et al. explicitly leave this open; no AcrIIA5–Cas9 co-structure exists.
• Why do the His66/Asn70/His73 mutants behave differently in vivo (active against SpCas9,
  per Garcia et al.) and in vitro (inactive, per An et al.)?
• Whether the in vivo sgRNA cleavage is a consequence of destabilised Cas9–sgRNA assembly
  exposing the guide to host RNases, as An et al. propose, has not been tested directly.